Github user StephanEwen commented on the pull request:
https://github.com/apache/flink/pull/1831#issuecomment-200951040
Thanks for describing this.
This has quite some big implications, as far as I can see it. The state in
the connected stream is now a "broadcast state" not partitioned, so allowing to
do that on the key/value state probably breaks some ongoing efforts, like
scaling, etc.
How about a more clean separation of these things:
1. The connected streams (fully partitioned or not)
2. Broadcast inputs, which are similar to the "side inputs" in cloud
dataflow or the broadcast variables in DataSet.
That gives us
- clean semantics, behavior that users can work with
- We do not overcomplicate the key/value abstraction
- and we can also make sure we checkpoint the broadcast state once only
(rather than on each parallel subtask).
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
